On the relationship between collagen- and carbonate-derived carbon isotopes with implications for the inference of carnivore dietary behavior

Studies of Rancho La Brea predators have yielded disparate dietary interpretations when analyzing bone collagen vs. enamel carbonate—requiring a better understanding of the relationship between stable carbon isotopes in these tissues. Stable carbon isotope spacing between collagen and carbonate (Δ_ca-co) has also been used as a proxy for inferring the trophic level of mammals, with higher Δ_ca-co values indicative of high carbohydrate consumption. To clarify the stable isotope ecology of carnivorans, past and present, we analyzed bone collagen (carbon and nitrogen) and enamel carbonate (carbon) of extinct and extant North American felids and canids, including dire wolves, sabertooth cats, coyotes, and pumas, supplementing these with data from African wild dogs and African lions. Our results reveal that Δ_ca-co values are positively related to enamel carbonate values in secondary consumers and are less predictive of trophic level. Results indicate that the foraging habitat and diet of prey affects Δ_ca-co in carnivores, like herbivores. Average Δ_ca-co values in Pleistocene canids (8.7+/−1‰) and felids (7.0+/−0.7‰) overlap with previously documented extant herbivore Δ_ca-co values suggesting that trophic level estimates may be relative to herbivore Δ_ca-co values in each ecosystem and not directly comparable between disparate ecosystems. Physiological differences between felids and canids, ontogenetic dietary differences, and diagenesis at Rancho La Brea do not appear to be primary drivers of Δ_ca-co offsets. Environmental influences affecting protein and fat consumption in prey and subsequently by predators, and nutrient routing to tissues may instead be driving Δ_ca-co offsets in extant and extinct mammals.

Introduction

Rancho La Brea (RLB) is one of the best preserved and studied fossil localities in the world, providing insight into the ecology of ancient mammals over the past ~50,000 years in southern California (e.g., Merriam and Stock, 1932; Stock and Harris, 1992; Van Valkenburgh and Hertel, 1993; DeSantis et al., 2012, 2019; Meachen and Samuels, 2012). Herbivorous mammals trapped in asphalt seeps (“pits”) attracted carnivores that also became trapped, resulting in one of the world’s best localities for studying ancient predators and their prey (Stock and Harris, 1992). As predators normally occur in lower numbers than their prey in modern ecosystems, they are also typically rare in fossil accumulations. However, because of their abundance in these deposits, Rancho La Brea provides unique insight into predator and prey dynamics in the Late Pleistocene—a time characterized by dramatic climatic change and the arrival of humans in North America (Stock and Harris, 1992; Barnosky et al., 2004). One productive avenue of research has been the inference of diet based on stable isotopes, allowing deep insight into the ecosystem dynamics of Late Pleistocene Rancho La Brea (e.g., Coltrain et al., 2004). However, recent work on isotopic values from RLB carnivores has revealed a clear discrepancy between values derived from collagen vs. enamel carbonate (DeSantis et al., 2019, 2020), complicating dietary inferences from either tissue. The goals of this study are to quantify and test potential drivers responsible for collagen/carbonate offsets generally, and to use these understandings to clarify dietary inference at RLB.

Stable isotopes and dietary inference

Stable isotope analyses of bone and/or tooth enamel have clarified the dietary ecology of numerous animals across the globe, past and present (e.g., Cerling et al., 1997; DeSantis et al., 2009; Secord et al., 2012). While studies of extant animals benefit from the ability to sample different tissues (including hair, feathers, scales, blood, etc.), archeological and paleontological studies of ancient life are limited to the analysis of tissues that preserve original biogenic stable isotope signatures in the fossil record (e.g., bones and teeth). Enamel is primarily inorganic (Teruel et al., 2015; Kendall et al., 2018) and stable isotopes of carbon and oxygen have been used to infer diet and climate, respectively, through time (e.g., DeNiro and Epstein, 1978; Cerling et al., 1997; Levin et al., 2006; DeSantis et al., 2009; Secord et al., 2012). Bone and dentin collagen, which are organic, also provide insights into the diets of ancient animals (e.g., Schoeninger et al., 1983; Bocherens et al., 1994; Finucane et al., 2006; Fuller et al., 2006; Kellner and Schoeninger, 2007; Lee-Thorp, 2008; Tung et al., 2016). Specifically, carbon isotope values measured in collagen (bone or dentin) provide information about the protein component of an animal’s diet (Ambrose and Norr, 1993) and nitrogen isotope values clarify trophic level with higher values indicative of feeding at higher trophic levels and/or more meat or fish consumption (Schoeninger et al., 1983). In archeological studies, both tooth enamel (or bone apatite) and bone collagen are analyzed to infer differences between whole diet (e.g., carbohydrates, fat, and protein) and protein components of diet, respectively (e.g., Ambrose and Norr, 1993). However, paleontological studies often employ only one tool, due to a lack of preserved collagen, interest in only isotopes associated with a particular tissue (e.g., δ¹⁸O in carbonate, δ¹⁵N in collagen; Bocherens et al., 1994; DeSantis et al., 2009), and/or the need to minimize destructive sampling. Therefore, direct comparisons between data types from the same specimens are rare.

Those studies that have examined both tooth enamel and bone collagen stable isotopes from the same individual specimens (e.g., Krueger and Sullivan, 1984; Lee-Thorp et al., 1989; Ambrose and Norr, 1993; Clementz et al., 2007, 2009; Murphy et al., 2007; Bocherens et al., 2017; Codron et al., 2018) indicate that isotope values from the different tissues record different aspects of their diet. There is broad consensus that δ¹³C_collagen reflects the protein component of diet, while δ¹³C_carbonate reflects whole diet (e.g., protein, carbohydrates, and fats; Ambrose and Norr, 1993) and that the isotopic spacing between carbonate and collagen (Δ_ca-co, δ¹³C_carbonate - δ¹³C_collagen) varies depending on trophic category (Krueger and Sullivan, 1984; Lee-Thorp et al., 1989; Clementz et al., 2009). Specifically, the lowest Δ_ca-co values occur in carnivores (averaging 4.3+/−1.0‰ to 4.8+/−0.4‰; Lee-Thorp et al., 1989; Clementz et al., 2009) while herbivores have higher Δ_ca-covalues (6.8+/−1.4‰ to 7.6+/−0.5‰; Lee-Thorp et al., 1989; Clementz et al., 2009). Animals with lower Δ_ca-co values are often attributed as having more meat and/or fat in their diet than animals with higher Δ_ca-co values—Δ_ca-co values are therefore used to infer degree of meat and/or fat consumption in ancient mammals (Lee-Thorp et al., 1989; Clementz et al., 2009). In Canis lupus, Δ_ca-co increases with increased δ¹³C_carbonate values (p = 0.016, R = 0.51; per our analysis of the supplemental data in Clementz et al., 2009); however, there is no predictable relationship between Δ_ca-co in bone apatite or bone collagen δ¹³C values in African carnivores (Codron et al., 2018). Further, the significant positive-relationship between Δ_ca-co values and δ¹³C_carbonate values in a sample of dire wolves (Aenocyon dirus, previously known as Canis dirus) at Rancho La Brea is similar to the relationship between Δ_ca-co and δ¹³C_collagen values in the African mammals included in Codron et al. (2018; DeSantis et al., 2020, see Figure 1B).

FIGURE 1

Figure 1. Stable carbon isotope data (collagen and carbonate) for all fossil specimens examined from Rancho La Brea (n = 71). Mean values are denoted in bold (bars indicate 1 standard deviation, n-1). All collagen data are from bone collagen and carbonate data are from tooth enamel (lower first molars), with the exception of A. dirus from Locality 3,874 which are from incisor dentin collagen and incisor enamel, respectively. The felid S. fatalis is noted in red, with all canids noted in blue; the same shade of light blue is used for canids from pits 61/67. The scales on the x-axis are of the same magnitude for the two tissues.

Offset values between carbonate and collagen can be highly variable among herbivores, specifically African mammals (primarily ungulates, especially bovids) and Australian kangaroos (Murphy et al., 2007; Codron et al., 2018). Murphy et al. (2007) demonstrated a significant negative relationship between Δ_ca-co in herbivorous kangaroos and aridity (i.e., water availability index; higher Δ_ca-co values when less water is available). Codron et al. (2018) instead documented higher Δ_ca-co values with increased δ¹³C_collagen values in African mammals, which they attribute to both physiological differences between ruminants and non-ruminants and dietary differences between grass and browse (e.g., effects of higher amounts of protein in ¹³C deplete browse and higher amounts of CH₄ produced when consuming grass). In closely related ruminant taxa (i.e., members of the family Bovidae), Δ_ca-co values calculated from teeth (enamel carbonate and dentin collagen) span the full range of values from 3.6 to 14.8‰; the second lowest offset value (4.3‰) and highest offset value (14.8‰) also occur with the same species—springbok (Antidorcas marsupialis). Further, when re-analyzing and comparing just tooth data (enamel carbonate and dentin collagen) from Codron et al. (2018) from ruminants in the family Bovidae and hindgut fermenters (in the families Equidae and Rhinocerotidae), the slopes and R² values are indistinguishable (slopes are 0.21 and 0.20, R² values are 0.41 and 0.43, respectively). Thus, it is unlikely that ruminant or non-ruminant physiology is the primary driver of Δ_ca-co values; instead, the proportion of grass and browse may be a larger driver of herbivore Δ_ca-co values. Specifically, higher amounts of protein enriched browse with ¹³C deplete values yield both lower δ¹³C carbonate values and smaller Δ_ca-co values, while protein deplete C₄ grasses also result in more methane production (that is isotopically depleted in ¹³C, leaving behind more ¹³C enriched carbon from which tissues are synthesized, in both ruminants and hind-gut fermenters) and are correlated with higher δ¹³C carbonate values and higher Δ_ca-co values. Discrepancies between how herbivore and carnivore Δ_ca-co values correlate with stable isotope values of collagen and enamel are apparent and more work on carbonate-collagen relationships is needed to elucidate potential drivers of Δ_ca-co values and the efficacy of using Δ_ca-co values to infer ecological information such as trophic level in extinct and extant mammals.

Stable isotopes and dietary inferences in RLB carnivores

Despite extensive study of Rancho La Brea carnivores, there are differences between the dietary interpretations of Late Pleistocene mammals at Rancho La Brea as inferred from bone collagen versus enamel carbonate stable isotope values (Coltrain et al., 2004; Feranec et al., 2009; Fuller et al., 2014, 2020; DeSantis et al., 2019, 2020; Van Valkenburgh et al., 2020). Seminal work analyzed the carbon and nitrogen isotope ratios of bone collagen from Rancho La Brea carnivores and concluded that sampled individuals of Smilodon fatalis (the sabertooth cat) and Aenocyon dirus (the dire wolf, previously known as Canis dirus) had indistinguishable carbon and nitrogen isotope values and were therefore competing for similar prey (Coltrain et al., 2004). This idea persisted until dietary interpretations based on tooth enamel carbonate indicated the consumption of different prey (DeSantis et al., 2019, 2020). Specifically, DeSantis et al. (2019, 2020) found that A. dirus yielded statistically greater δ¹³C_carbonate values than S. fatalis—which was interpreted as S. fatalis primarily having consumed prey that occupied denser forests/shrubland ecosystems, while A. dirus consumed prey from more open grassland environments. The reasons for this discrepancy between the two tissues remain unclear (Fuller et al., 2014, 2020; DeSantis et al., 2019, 2020), and have led some to question the recent dietary interpretations of Rancho La Brea carnivorans (Van Valkenburgh et al., 2020). While DeSantis et al. (2019, 2020) speculate on some of these discrepancies, they concentrated on an enamel-based interpretation of diet.

To clarify the paleobiology of Rancho La Brea carnivores and resolve relationships between enamel carbonate and bone collagen in both extinct and extant mammals, tooth enamel and bone collagen were examined in extant canids and felids from C₃ and C₄ ecosystems in North America and Africa, in addition to Rancho La Brea carnivorans. Specifically, we examined Late Pleistocene specimens from Rancho La Brea (A. dirus, C. latrans, S. fatalis), modern specimens from predominately C₃ ecosystems in southern California (C. latrans and Puma concolor), and modern specimens from predominately C₄ ecosystems throughout Africa (Lycaon pictus and Panthera leo). We test the hypotheses that Δ_ca-co values in carnivorous mammals (1) are largely consistent and independent of either δ¹³C_carbonate or δ¹³C_collagen values (~4.3+/−1.0‰ to 4.8+/−0.4‰; Lee-Thorp et al., 1989; Clementz et al., 2009), and (2) are indicative of trophic level. We also further clarify relationships between carbonate (enamel, bone, and dentin) and collagen (bone and dentin) in extant African canids and felids.

Statistical considerations of isotopic data from different tissues

We recognize that the correlations between Δ_ca-co values and δ¹³C_carbonate or δ¹³C_collagen, may, in part, be influenced by plotting these values against each other; but these plots and correlations permit comparison to other studies, and provide information regarding the source of these isotopic data. As Auerswald et al. (2010) demonstrated, if pairs of values are drawn randomly from populations of variables A and B of equal variance, the difference A-B will be biased to high values when A is high and to low values when A is low. Hence an A-B vs. A correlation with a slope of +1 is generated; conversely A-B vs. B will yield a slope of-1. In the unrealistic scenario that δ¹³C variations in the two tissues were completely independent (i.e., if their responses to environmental and physiological changes were totally uncoupled) the slopes of these relationships would approach +1 and-1 respectively, whereas if the isotopic responses of the two tissues were identical the slopes would both be zero. Since Δ_ca-co is defined as δ¹³C_carbonate − δ¹³C_collagen, all other things being equal one expects to see a positive correlation between Δ_ca-co and δ¹³C_carbonate and a negative correlation between Δ_ca-co and δ¹³C_collagen. Additionally, if these relationships were a product of random sampling, we would not expect δ¹³C values of different tissues to yield different relationships with Δ_ca-co values. See Caut et al. (2010) for a detailed reply to Auerswald et al. (2010) and discussion regarding the rationale for these sorts of comparisons that are prevalent in paleobiology and archeology (e.g., Krueger and Sullivan, 1984; Lee-Thorp et al., 1989; Ambrose and Norr, 1993; Murphy et al., 2007; Clementz et al., 2009; Codron et al., 2018).

Differences in isotopic variance of tissue types are evident from our data set and the extensive data sets of Murphy et al. (2007) and Codron et al. (2018). However, this is not necessarily an intrinsic property of these tissues; studies have shown that δ¹³C_collagen ranges are greater than those in δ¹³C_carbonate in humans when their diets include protein sources with disparate δ¹³C values (e.g., marine protein enriched in ¹³C in addition to terrestrial C₃ resources that are ¹³C deplete; Tung et al., 2020).

Materials and methods

To decipher the conflicting data interpretations between the stable isotope values of bone collagen and tooth enamel carbonate in Rancho La Brea carnivores, jawbone specimens containing teeth (n = 71) were obtained for stable isotope analysis of both bone collagen and tooth enamel carbonate in the same specimen. Lower first molars were sampled in all taxa due to the larger size of these carnassial teeth and the need to minimize specimen destruction. In the case of Aenocyon dirus from the University of California Museum of Paleontology (UCMP) Locality 3,874, enamel was from incisors while collagen was from dentin (both near the enamel/dentin junction and down through the root, due to the need for appropriately large amounts of this tissue, as these teeth were isolated and lacked associated bone for bone collagen analysis; as noted in Supplementary Table S1, Supplemental Tables 1, 2). Rancho La Brea specimens comprised Aenocyon dirus (Pits 61/67, n = 21; Locality 3,874, n = 8), Canis latrans (Pits 61/67, n = 19; Pit 10, n = 12), and Smilodon fatalis (Pit 61/67, n = 11; Supplementary Table S1, Supplemental Table 1). Modern predators from southern California (n = 19; Canis latrans, Puma concolor) and from Africa (n = 18; Lycaon pictus, Panthera leo) were also analyzed for both bone collagen and tooth enamel (Supplementary Tables S1, S2, Supplemental Table 2). The African specimens were also analyzed for bone (carbonate) and dentin (carbonate and collagen), in addition to enamel (carbonate) and bone (collagen). When more than one analysis was run on the same tissue, the sample was first homogenized and then split up after drilling to reduce temporal variability that could otherwise occur if tissues were not homogenized after sampling. While we would have liked to have sampled dentin carbonate and collagen from all Rancho La Brea specimens, this was not possible due to the scale and scope of this study and need to minimize specimen destruction on the fossil specimens (collagen analyses require significantly more sample as compared to carbonate). As tooth enamel carbonate and bone collagen are the most commonly analyzed tissues at Rancho La Brea and other Late Pleistocene fossil sites, we concentrated on these tissues for the bulk of our analyses. The addition of dentin carbonate and collagen and bone carbonate of modern African specimens can provide insights that can be followed up with future studies. Rancho La Brea specimens are housed at the University of California, Berkeley (UCMP Locality 3,874) and the La Brea Tar Pits and Museum (LACMHC). Modern specimens from southern California are from the Santa Barbara Museum of Natural History (SBMNH) and modern specimens from Africa are from the American Museum of Natural History (Mammalogy Collections; AMNH).

Carbonate samples from both teeth and bone were drilled with a low-speed dental-style drill and carbide dental burrs (<1 mm burr width). The resulting carbonate powder samples (enamel, dentin, bone) were pretreated with 30% hydrogen peroxide for 24 h, and 0.1 N acetic acid for 18 h, to remove organics and secondary carbonates, respectively (Koch et al., 1997; DeSantis et al., 2009). Sample preparation followed procedures outlined in DeSantis et al. (2019) and included the rinsing of samples with ultra-pure water and subsequently centrifuging them between all treatment steps. Approximately 1 mg of each sample was then analyzed on a VG Prism stable isotope ratio mass spectrometer with an in-line ISOCARB automatic sampler in the Department of Geological Sciences at the University of Florida. All bone or dentin samples analyzed for collagen carbon or nitrogen isotope ratios were cut with a Dremel® UltraSaw ™ cutting wheel or a diamond coring drill (dentin samples also sampled for carbonate were subsequently drilled, i.e., a subset of the sample was removed for carbonate analysis).

Clean collagen was recovered from the asphalt-impregnated bones using a modified procedure based on the protocol developed by Fuller et al. (2014). Samples of crushed bone were sonicated in a solvent sequence of increasing polarity to remove bulk asphalt, demineralized overnight with HCl, rinsed with ultrapure water, and gelatinized overnight at pH 2, at an elevated temperature (75°C) sufficient to break up collagen molecules into fragments of a molecular weight on the order of 10’s of kDa, while leaving massive residual asphalt aggregates intact. The gelatin solution was then purified using two sequential ultrafiltrations to select an intermediate molecular weight fraction (5–100 kDa) which was frozen with liquid nitrogen and lyophilized overnight. Elemental analyses (‰C and ‰N) and stable isotope ratios were determined by flash combustion of ~0.7 mg samples of collagen in tin capsules, and analysis of the gas products using a Fisons NA1500NC elemental analyzer/Finnigan Delta Plus isotope ratio mass spectrometer combination at the Keck AMS laboratory at University of California Irvine. Stable isotope data are reported in conventional delta (δ) notation for carbon (δ¹³C), oxygen (δ¹⁸O), and nitrogen (δ¹⁵N), where δ¹³C (parts per thousand, ‰) = [(R_sample/R_standard)−1]*1,000, and R = ¹³C/¹²C; and, δ¹⁵N (parts per mil, ‰) = ((R_sample/R_standard)-1)*1,000, and R = ¹⁸O/¹⁶O, R = ¹⁵N/¹⁴N on the VPDB (for carbon and oxygen; Pee Dee Belemnite, Vienna Convention; Coplen, 1994) and AIR (for nitrogen; Mariotti, 1983) scales. The analytical precision for carbonate δ¹³C and δ¹⁸O is ±0.1‰, based on replicate analyses of samples and standards (NBS-19). The analytical precisions for δ¹³C and δ¹⁵N from bone collagen are ±0.1‰, and ±0.2‰ respectively, based on replicate analyses of samples and standards.

Studies utilizing carbonate-collagen spacing use “capital delta notation” values, calculated as follows: Δ_ca-co = δ¹³C_carbonate − δ¹³C_collagen (e.g., Krueger and Sullivan, 1984; Lee-Thorp et al., 1989; Ambrose and Norr, 1993; Clementz et al., 2007; Murphy et al., 2007; Clementz et al., 2009; Bocherens et al., 2017; Codron et al., 2018). While we use these calculations to be directly comparable with prior analyses, we also calculated the fractionation factor (α) and isotopic enrichment (ε; Supplemental Table 1) similar to Passey et al. (2005), based on Friedman and O’Neil (1977). Specifically, α_{carbonate-collagen} = (1,000+ δ_carbonate)/ (1,000+ δ_collagen), and ε_{carbonate-collagen} = (α_{carbonate-collagen} – 1)1000 (Friedman and O’Neil, 1977). Differences between Δ_ca-co and ε_{carbonate-collagen} are small within the ~10‰ range here noted (Cerling and Harris, 1999); of all specimens examined here, the average difference is 0.133‰ +/−0.0518 SD, n-1 (ranging from 0.029 to 0.207‰). These differences are minor (approximately the error of the mass spectrometers), do not significantly affect the patterns observed, and the use of Δ_ca-co is important for consistency with published works (e.g., Krueger and Sullivan, 1984; Lee-Thorp et al., 1989; Ambrose and Norr, 1993; Murphy et al., 2007; Clementz et al., 2009; Codron et al., 2018); however, we suggest that α and ε be reported and discussed moving forward, as noted by Passey et al. (2005; all values are reported in the Supplemental information).

Parametric statistical tests were employed to compare stable isotope values among taxa, between tissues, and assess relationships between isotopic values and Δ_ca-co values (e.g., Student’s t-tests, including paired t-tests when comparing tissues in the same individuals, linear regressions). No corrections (including the Suess correction) were made to reported and analyzed data (Supplementary Tables S1, S2, Supplemental Tables 1, 2). While we recognize that δ¹³C values today are approximately 1.5‰ lower than during the Late Pleistocene (Cerling et al., 1997; Cerling and Harris, 1999; Passey et al., 2005), we are focused on assessing relationships between δ¹³C values from disparate tissues within individuals (e.g., relationships between δ¹³C_carbonate and Δ_ca-co, as opposed to comparing diets of extant and extinct predators) and the majority of comparisons are done within modern or within Late Pleistocene samples. Further, it is possible that the δ¹³C values of the atmosphere (which are a function of CO₂ levels; Keeling, 1979; Keeling et al. 1979) affect how δ¹³C values are incorporated into various tissues via plants and prey, as we suggest in the discussion. All things being equal, plants exposed to higher atmospheric CO₂ levels have lower nutrients and lower nitrogen content/lower protein content but more carbon per unit biomass, and C₄ grasses have lower nutritional quality (Conroy, 1992; Barbehenn et al., 2004). This can impact nutrient routing to tissues in consumers and have downstream impacts on carnivores. Further, all studies to date have examined relationships between δ¹³C_carbonate and Δ_ca-co in modern taxa with no Suess corrections applied (e.g., Lee-Thorp et al., 1989; Ambrose and Norr, 1993; Murphy et al., 2007; Clementz et al., 2009; Codron et al., 2018), and we aimed to produce data directly comparable to those data sets.

Results

Bone collagen and enamel carbonate δ¹³C values from Rancho La Brea

Stable carbon isotopes measured in bone collagen of S. fatalis (Pits 61/67) are indistinguishable from A. dirus (Pits 61/67; p = 0.141; Figure 1), consistent with prior work (Coltrain et al., 2004; Fuller et al., 2014, 2020). In contrast, stable carbon isotopes measured in enamel carbonate of S. fatalis (Pits 61/67) are significantly lower than in A. dirus (Pits 61/67; p < 0.0001; Figure 1), which is also consistent with prior analyses of enamel carbonate (DeSantis et al., 2019, 2020). When comparing the three most abundant carnivores at Rancho La Brea from the same deposit (A. dirus, C. latrans, and S. fatalis from Pits 61/67), bone collagen carbon isotope values are indistinguishable (p = 0.187), while enamel carbon isotope values of the canids A. dirus and C. latrans are significantly greater than S. fatalis (p = 0.0002, p = 0.004, respectively). Further, stable nitrogen isotopes ratios measured in the morphologically inferred hypercarnivores A. dirus and S. fatalis are significantly greater than C. latrans (p < 0.0001 for both), while A. dirus and S. fatalis are indistinguishable from one another (p = 0.332). Despite the fact that bone collagen and enamel carbonate integrate information about different portions of animal diets (protein component versus whole diet; Ambrose and Norr, 1993), there is a significant positive relationship between enamel δ¹³C_carbonate and δ¹³C_collagen from bone in both the fossil Rancho La Brea specimens (p < 0.0001, R² = 0.25, R = 0.5 in Rancho La Brea taxa, Figure 2A) and modern southern California (p = 0.0002, R² = 0.57, R = 0.75; Supplemental Table 1) specimens examined here.

FIGURE 2

Figure 2. Relationship between δ¹³C_carbonate and δ¹³C_collagen values (A), and Δ_ca-co and δ¹³C_collagen values (B), δ¹⁵N_collagen values (C), and δ¹³C_carbonate values (D; n = 108). The regression lines and R² values correspond to text in the same color and are also noted in Supplementary Table S3. Note, the modern C₃ regression lines in (A) and (B) includes modern North American taxa and one African wild dog with a C₃ value (Supplementary Tables S3, S4 only include geographically similar taxa). The legend in (B) applies to all (A–D). A light-gray dashed-line with a slope of 1 is present in the background of each panel for reference. *Denotes significant relationships (p < 0.05).

Collagen-carbonate δ¹³C spacing in carnivorans from southern California

There is a strong and significant positive relationship between Δ_ca-co and δ¹³C_carbonate in enamel from both extant and extinct southern California canids and felids (R² = 0.88, R = 0.94; p < 0.0001; Figure 2D; Supplementary Table S3). This relationship additionally holds true when modern and fossil specimens are evaluated separately (modern: R² = 0.58, R = 0.76, p = 0.0002; fossil: R² = 0.72, R = 0.85, p < 0.0001). Because modern specimens of the canid C. latrans and the felid P. concolor have lower δ¹³C_carbonate values than the extinct felids and canids from Rancho La Brea, likely due to their consumption of prey from ecosystems with increased canopy cover or denser shrubs (Feranec and DeSantis, 2014; DeSantis et al., 2019) and the Suess effect that accounts for ~1.5‰ lower δ¹³C_carbonate values (Cerling et al., 1997; Cerling and Harris, 1999; Passey et al., 2005); we graphically present the results of linear regressions for fossil and modern specimens separately (see Materials and Methods, Figure 2). However, all relationships between enamel δ¹³C_carbonate and Δ_ca-co were examined and reported (Supplementary Table S3). Δ_ca-co values and δ¹³C_collagen values exhibit no significant correlation when including either Rancho La Brea specimens (R = 0.03, p = 0.767) or modern specimens from southern California (R = 0.14, p = 0.531); however, there is a weak but significant positive relationship when both Rancho La Brea specimens and modern C₃ specimens are included (Figure 2B; Supplementary Table S4). In contrast to δ¹³C_collagen values, δ¹³C_carbonate values can be used to predict Δ_ca-co values; thus, δ¹³C_collagen values can be estimated within 0.5‰.

Rancho La Brea specimens have average Δ_ca-co values of 8.5+/−1.2‰ (ranging from 6 to 11‰, n = 71) and modern specimens from southern California have average Δ_ca-co values of 4.6+/−1‰ (ranging from 3.2 to 7‰, n = 19). Rancho La Brea canid and felid Δ_ca-co values (8.7+/−1.0‰, 7.0+/−0.7‰, respectively) overlap with modern herbivore Δ_ca-co values (Lee-Thorp et al., 1989; Murphy et al., 2007; Clementz et al., 2009; Codron et al., 2018) and are higher than modern carnivores based on the literature (averaging 4.3+/−1.0‰ to 4.8+/−0.4‰; Lee-Thorp et al., 1989; Clementz et al., 2009).

Most notably, Δ_ca-co values of both Rancho La Brea carnivores and modern carnivores from southern California are not significantly (negatively) correlated with δ¹⁵N_collagen values as was expected (Figure 3C; Supplementary Table S4), based on prior research that documents higher Δ_ca-co values in more omnivorous or herbivorous taxa (e.g., Lee-Thorp et al., 1989; Clementz et al., 2009). C. latrans from Pits 61/67 at Rancho La Brea exhibit significantly lower δ¹⁵N_collagen values than both A. dirus (p < 0.0001) and S. fatalis (p < 0.0001) from the same location, suggesting that coyotes were more omnivorous than the larger, more abundant, and inferred hypercarnivores (A. dirus and S. fatalis; Merriam and Stock, 1932; Stock and Harris, 1992). Yet, the mean Δ_ca-co value of C. latrans is indistinguishable from A. dirus (p = 0.411, Mann–Whitney U).

FIGURE 3

Figure 3. Relationships between δ¹³C_carbonate (bone, dentin, or enamel) and Δ_ca-co (calculated from carbonate-collagen from different tissues) in African herbivores (A) and carnivores (A,B). Data from herbivores are from Codron et al. (2018), in addition to some of the carnivore bone samples noted in panel A. The regression lines and R² values correspond to text in the same color and are also noted in Supplementary Table S3. The dashed red-line in (B) is for enamel carbonate-bone collagen (squares). *Denotes significant relationships (p < 0.05).

Collagen-carbonate δ¹³C spacing and tissue specific enrichment in carnivorans from Africa

As in southern California, there is a strong (R² = 0.65, R = 0.81) and significant positive relationship (p = 0.0001; Supplementary Table S3) between δ¹³C_carbonate in enamel and Δ_ca-co in African canids and felids from predominantly C₄ ecosystems. One individual with a δ¹³C_{enamel carbonate} value of −12.3‰, interpreted as consuming significant C₃ prey (and ~ 3 standard deviations from the mean) fell on the regression line from southern California extant and extinct taxa (Figure 2A) and was not included in the analysis of exclusively African (C₄) carnivores; however, it was included in the regression of C₃ taxa in Figures 2B,D, and relevant regressions of all carnivores (or modern carnivores; per Materials and Methods). δ¹³C_collagen and Δ_ca-co values are not significantly correlated to one another in exclusively C₄ carnivorans (R² = 0.13, p = 0.158; Figure 2B; Supplementary Table S4), while δ¹⁵N_collagen and Δ_ca-co values exhibit a significant positive relationship (R² = 0.31, p = 0.021; Supplementary Table S4, all data located in Supplemental Table 2).

There are differences among stable carbon isotope values measured in different tissues (Supplementary Table S5). Collagen consistently has lower δ¹³C values than carbonate samples by 3.6–5.5‰, which is consistent with prior work (e.g., Lee-Thorp et al., 1989; Ambrose and Norr, 1993). Enamel δ¹³C_carbonate values are significantly higher than both dentin δ¹³C_carbonate values (p < 0.001) and bone δ¹³C_carbonate values (p = 0.004), though mean differences are low (1.7 and 1.1‰, respectively; Supplementary Table S5). Further, dentin δ¹³C_carbonate values are significantly lower than bone δ¹³C_carbonate values (p = 0.029; Supplementary Table S5). All carbon isotopes signatures from all tissues sampled (enamel, dentin, and bone) yield significantly different values from one another (all p-values≤0.001; Supplementary Table S5). Only dentin and bone δ¹³C_collagen values yield mean values indistinguishable from one another (p = 0.504; Supplementary Table S5). Mean δ¹⁵N_collagen values from dentin are significantly higher than bone collagen (p < 0.001; Supplementary Table S5); however, the mean difference is quite small (0.9 +/−0.7‰; Supplementary Table S5). Further, the tissues with the strongest correlations between δ¹³C values (dentin carbonate-dentin collagen, and bone carbonate-bone collagen) exhibit no significant relationships between apatite (dentin or bone) δ¹³C_carbonate values and Δ_ca-co in African carnivores (Supplementary Table S4). Enamel δ¹³C_carbonate values are consistently positively correlated with Δ_ca-co values (whether Δ_ca-co is based on dentin or bone collagen; Supplementary Table S3). Dentin δ¹³C_carbonate values are also significantly related to Δ_ca-co values when bone collagen is used; however, neither bone nor dentin exhibit any significant relationships between carbonate δ¹³C values and Δ_ca-co values when calculated from just bone or just dentin tissues (carbonate/collagen; Figure 3). Further, African carnivores from C₄ ecosystems have average Δ_ca-co values of 5.5+/−1.2‰ when comparing enamel and bone, however Δ_ca-co values are lower when bone carbonate and bone collagen are used (4.4+/−0.6‰ for C₄ carnivores, 4.7+/−0.6‰ for Lycaon pictus, and 4.2+/−0.5‰ for Panthera leo) and closer to values reported in the literature (averaging 4.3+/−1.0‰ to 4.8+/−0.4‰; Lee-Thorp et al., 1989; Clementz et al., 2009).

Collagen-carbonate δ¹³C and δ¹⁸O variances in carnivorans

Carbonate and collagen tissues do vary to different extents, with δ¹³C_carbonate values having greater variances than δ¹³C_collagen values (p < 0.0001 for both Levene’s Median test and Bartlett’s test) and a standard deviation almost double that of collagen (1.3 vs. 0.7‰) in Rancho La Brea samples. Modern specimens from C₃ and C₄ ecosystems were equivocal (C₃ specimens, p = 0.111 and p = 0.039 for Levene’s Median test and Bartlett’s test; C₄ specimens, p = 0.025 and p = 0.081 for Levene’s Median test and Bartlett’s test, respectively), but sample sizes were significantly lower (n = 20, n = 17, respectively) than Rancho La Brea specimens (n = 71).

Stable oxygen isotope ratios were measured and compared among all carbonate samples from southern California to assess if Rancho La Brea fossils had anomalously low δ¹⁸O values, potentially due to diagenesis. Modern canid and felid (i.e., Canis latrans and Puma concolor, respectively) enamel δ¹⁸O values ranged from-5.0 to-0.6 (mean = −2.6+/−1.3‰, n = 19) and were indistinguishable in both mean value (p = 0.336) and variance (p = 0.955) from δ¹⁸O values of fossil specimens that include C. latrans, A. dirus, and S. fatalis from Rancho La Brea (δ¹⁸O values ranged from-5.3 to 2.3; mean = −2.3+/−1.3‰, n = 71).

Discussion

Identifying the primary drivers of collagen-carbonate spacing in carnivores requires the testing of numerous hypotheses. Below we evaluate hypotheses pertaining to ontogenetic dietary differences, diagenesis, and physiological differences between canids and felids, eliminating these as potential drivers with data provided from this paper and published studies. We are unable to eliminate hypotheses pertaining to tissue and environmental/dietary differences, and their downstream impacts on Δ_ca-co spacing. Further work is needed to clarify the exact mechanism(s) responsible for Δ_ca-co spacing (including the role of trophic level in different ecosystems); regardless, this work helps to advance our understanding of a challenging problem and cautions the use of static offset values between collagen and carbonates in future work (as has become a more common practice; e.g., Clementz et al., 2009; Smith et al., 2022).

The absence of significant ontogenetic dietary differences and diagenetic effects on collagen-carbonate δ¹³C spacing

Tissues such as bone collagen and enamel carbonate mineralize at different times and undergo significant to no remodeling, respectively. While it is possible that ontogenetic differences in diet are responsible for the isotopic differences between bone collagen (representing a later in life signal due to continuous remodeling) and enamel carbonate (occurring earlier in life when the teeth mineralized and representing a discrete interval of time) in the carnivores examined, this would require dramatic dietary changes over time that are possible but not probable. The canids studied would need to eat prey from more open grassland environments as young individuals and then shift to eating forest-dwelling prey as adults, in a manner that predictably correlates with δ¹³C_carbonate values in all felids and canids studied. This depiction of canid and felid ecology also does not align with morphological interpretations of extinct taxa or our knowledge of extant felid and canid behavior in the wild (Elliot et al., 1977; Bekoff and Wells, 1982). Specifically, the morphology of S. fatalis and A. dirus suggests terrestrial and cursorial hunting in these taxa, respectively (Samuels et al., 2013), while observational data suggests that extant felids typically use vegetation as cover to ambush prey while extant canids rarely utilize cover (Elliot et al., 1977; Bekoff and Wells, 1982). Additionally, both A. dirus and S. fatalis are interpreted as social predators (Carbone et al., 2009; McHorse et al., 2012); in extant social carnivores, juveniles typically consume the same prey as adults (i.e., meat and/or carcasses are shared, as observed in African lions; Schaller, 1972). Published adult δ¹³C_collagen values that span multiple pits (Fuller et al., 2014, 2020) prepared via ultrafiltration (as was done here), were indistinguishable from juvenile δ¹³C_collagen values from Locality 3,874 (p = 0.378; means of −18.4 and −18.5‰ respectively), indicating that juvenile dire wolves are eating similar prey as adults. Juvenile δ¹³C_collagen values are from tooth dentin in A. dirus (Supplemental Table 1) and bone collagen from one juvenile humerus (Fuller et al., 2020). While δ¹³C_collagen values from tooth dentin (laid down during the animal’s youth) in A. dirus (Locality 3,874; no published radiocarbon dates, a general locality name for material collected by the University of California Museum of Paleontology; Feranec et al., 2009) are significantly greater than δ¹³C_collagen values from bone collagen from A. dirus deposited in Pits 61/67 (~11,581 +/−3,768 years before present, calibrated; O’Keefe et al., 2009), these differences may be a result of different ecologies during different sampling periods (i.e., Locality 3,874 may represent a much broader, narrower, or different span of time than Pits 61/67; Feranec et al., 2009, O’Keefe et al., 2009). Further, extant African canids and felids reveal only minor differences in δ¹³C_carbonate values between bone (a more recent diet) and both enamel (1.1‰) or dentin (0.7‰; both enamel and dentin are laid down earlier in life; also see Figure 6 from Bocherens, 2015), suggesting that ‘whole diet’ ontogenetic differences are minor through time, with such minor differences potentially due to weaning. It should also be noted that δ¹³C_collagen are indistinguishable between dentin and bone collagen in extant African carnivores (Supplementary Table S5), indicative of similar dietary protein sources through ontogeny in African canids and felids.

A case could be made that the difference in variance at Rancho La Brea is due to bone turnover smoothing out differences that appear in δ¹³C values of the enamel, the latter of which is laid down over a relatively short period of the animal’s lifespan and is not subject to reworking (as is the case with bone; Hedges et al. 2007). However, this cannot explain the dire wolf data from Locality 3,874, where enamel δ¹³C is compared with collagen from dentin (which like enamel, does not turn over; Helfman and Bada, 1976) though it is of course possible that the sampling geometry of enamel and dentin represent slightly different times (separated by days to month, not years; Bocherens, 2015). Further, the tooth dentin analyzed from Locality 3,874 has significantly higher δ¹⁵N_collagen values than A. dirus from Pit 61/67 where bone collagen (not dentin collagen) was analyzed; suggesting that the dentin analyzed was laid down early in the life of the animal when it was still consuming a significant portion of its mother’s milk (i.e., 13.7+/−1.1‰ SD in dentin from Locality 3,874, 11.8 +/−0.5‰ in bone collagen from Pits 61/67; p < 0.0001). At locality 3,874, δ¹³C_carbonate values range from −12.6 to −7.0‰ while δ¹³C_collagen values of the same individuals range from −18.8 to −18‰, in the same teeth. Therefore, differences in δ¹³C variability likely stems from difference in δ¹³C variability of protein and whole diet sources, and not the timing or turnover of tissues. As mentioned above, lipids can be extremely ¹³C depleted relative to protein (Ramsay and Hobson, 1991) and may contribute to larger δ¹³C_carbonate ranges as compared to δ¹³C_collagen ranges. Hence, understanding relationships between δ¹³C values from different tissues and Δ_ca-co values derived from those tissues has unique challenges. This highlights the importance of experimental studies and the benefits of obtaining data from multiple tissue types (especially from fossils) when possible—both to best understand protein and whole diet sources while also providing further resolution to understanding relationships between δ¹³C and Δ_ca-co values.

If a nursing/weaning signal was apparent, we would expect both enamel and dentin to have significantly lower δ¹³C_carbonate values than bone δ¹³C_carbonate values (due to the consumption of fattier milk with lower δ¹³C values while nursing and/or weaning earlier in life; e.g. Richards et al., 2002; Fuller et al., 2006; Tsutaya and Yoneda, 2015), and this is not the case. Contributions from weaning are also less likely due to early weaning of canids (by ~5–8 weeks of age; Mech, 1974, Ewer, 1998, Nowak and Walker, 1999) while weaning occurs later in large felids (e.g., African lions, weaning by ~8–9 months; Haas et al., 2005). Further, studies that examined the effect of weaning on δ¹³C values in tissues, including plasma, hair, and tooth enamel in humans and other mammals, showed its effect was minimal (<1‰; Tsutaya and Yoneda, 2015). Studies of non-primate taxa, including herbivores and carnivores, showed no evidence of ¹³C enrichment in the plasma of offspring (0.0+/−0.6‰; Jenkins et al., 2001), while ¹³C enrichment of ~1‰ was observed in the tissues (hair, fingernails, and/or ribs) of exclusively breastfed human infants (Richards et al., 2002; Fuller et al., 2006). Further, serial sampling of the enamel of two adult upper canines from S. fatalis from Rancho La Brea (which developed and mineralized from near birth up to 25 months of age) did not exhibit a weaning signal or any trend in δ¹³C values over time (Feranec, 2004; Wysocki et al., 2015).

Other possible reasons for the difference between collagen and carbonate δ¹³C values include differences in tissue formation and/or post-mortem differences in diagenesis between tissues. Bone collagen is organic, while enamel is primarily inorganic (only ~6–10‰ organic content; Teruel et al., 2015; Kendall et al., 2018). It is possible that the bone collagen values were altered in some way that homogenized the isotopic values of the taxa examined, either in situ or during collection and cleaning (which included boiling in kerosene in the historical early twentieth century collections); or conversely that the larger spread in the enamel δ¹³C values (Figure 1) is due to variable diagenesis or contamination that affected the carbonate. However, the fact that Δ_ca-co values in modern taxa are predictable from δ¹³C_carbonate values, like those from Rancho La Brea (Figure 2A), suggests that neither of these processes is likely and that the origin of the greater variability of δ¹³C_carbonate values relative to δ¹³C_collagen is biogenic. Furthermore, isotopic difference in δ¹³C_collagen values is apparent between other taxa that co-occurred in Rancho La Brea (e.g., bison, horses, and camels as compared to A. dirus, C. latrans, and S. fatalis; Coltrain et al., 2004), again suggesting that bone collagen values were not homogenized due to taphonomic factors; and samples with C:N ratios of 3.1–3.4 (well within the range of 2.9–3.6 indicative of biogenic values; Ambrose et al., 1997) were included in this study. Similarly, δ¹³C_carbonate values of extinct taxa from Rancho La Brea are similar in both δ¹³C_carbonate values and in variability as other Pleistocene sites in North America (e.g., DeSantis et al., 2009, 2019, 2020; Feranec et al., 2009; Feranec and DeSantis, 2014; Jones and DeSantis, 2017). If the δ¹³C_carbonate values were diagenetically altered, we might expect anomalously low oxygen isotope values as oxygen isotopes are more susceptible to post-mortem alteration than δ¹³C values from structural carbonates from bioapatite (Wang and Cerling, 1994). However, oxygen isotopes values measured in enamel carbonate from fossil material are neither more variable nor anomalously lower than δ¹⁸O values from extant C. latrans and P. concolor from southern California (collected during the 20^th and 21^st centuries). Finally, a large set of published literature demonstrates that tooth enamel is inherently less prone to diagenetic alternation than other tissues like bone apatite, bone collagen, dentin apatite, and dentin collagen (e.g., Wang and Cerling, 1994; Collins et al., 2002; Koch, 2007; Lee-Thorp, 2008; MacFadden et al., 2010). Hence while diagenesis is possible at any fossil locality, several lines of evidence suggest that both types of tissues sampled in this study (i.e., enamel and bone) record biogenic isotope signals and that diagenesis is not a likely driver of the carbonate-collagen discrepancies we observe here.

Physiological similarities in collagen-carbonate δ¹³C spacing between canids and felids and the need for caution in making trophic level assessments

Δ_ca-co spacing has been previously used to identify differential dietary sources as well as trophic level, however data from this study suggest that such determinations should be made with caution. For example, larger values indicate when protein and carbohydrates come from different sources while smaller values indicate that both protein and carbohydrates are from similar sources. The work of Ambrose et al. (1997) documents that when ancient humans have Δ_ca-co spacing >4.4‰ they consumed C₄ carbohydrates and C₃ protein, while values under this threshold suggest that both carbohydrates and proteins originate from similar sources (e.g., marine protein). In herbivores, it is thought that consumption of C₄ vegetation and rumination contribute to higher Δ_ca-co values; ¹³C deplete methane produced via fermentation results in ¹³C enriched CO₂, which is subsequently incorporated into the body via blood bicarbonate (Hedges, 2003; Clementz et al., 2009; Codron et al., 2018). Thus, herbivorous mammals exhibit variability in Δ_ca-co values, even when consuming only primary productivity—with values ranging from 3.6 to 14.8‰ in African bovids, likely due to differences in the amounts of ¹³C depleted methane produced via digestion and subsequently removed from the remaining nutrient pool via expulsion of methane gas by the animal (Codron et al., 2018). Further, it should be noted that all bovids are ruminants, so differences in Δ_ca-co values likely have more to do with disparate diets than differences in digestive physiology—especially when one species (Antidorcas marsupialis) has Δ_ca-co values that range more than 10‰ (from 4.3 to 14.8‰; these wide ranging values may be related to stark differences in protein content in C₃ and C₄ plants consumed as discussed later). In hypercarnivorous mammals, fats, proteins, and carbohydrates (minimal to no consumption) should largely come from the same source, so that larger Δ_ca-co spacing would suggest individuals with more plant biomass in their diets (Clementz et al., 2009; Bocherens et al., 2017). It was presumed that higher lipid consumption by carnivores (especially hypercarnivores) was responsible for lower Δ_ca-co values (Krueger and Sullivan, 1984; Lee-Thorp et al., 1989; O’Connell and Hedges, 2017) and thus Δ_ca-co values would be predictive of trophic level (Clementz et al., 2009; Bocherens et al., 2017). However, the hypercarnivores examined here—including P. concolor, S. fatalis and A. dirus—have average offset values that range from 4.4+/−0.6 in P. concolor to 8.8+/−1.1 in A. dirus (when calculated from enamel carbonate and bone collagen) and are thus unrelated to trophic level (as inferred from δ¹⁵N_collagen values) in Rancho La Brea carnivores and extant carnivores from C₃ ecosystems (Supplementary Table S4). Contrary to expectations, extant hypercarnivores from C₄ ecosystems exhibit higher (not lower) Δ_ca-co values with higher δ¹⁵N_collagen values (Supplementary Table S4). Data from extant and extinct carnivores here suggest that the C₃ or C₄ environment the prey are foraging in has an impact on not only the δ¹³C_carbonate values, but also the offset between collagen and carbonate when Δ_ca-co is calculated from enamel carbonate and bone collagen (Figure 2)—though this may be related to fat consumption and/or digestive physiology via effects of vegetation on fat accumulation in prey or digestion of those prey in carnivores. Thus, trophic level estimations may be more nuanced and not directly comparable between disparate ecosystems.

Codron et al. (2018) documented a positive relationship between δ¹³C_carbonate and Δ_ca-co values in herbivores, combining teeth and bones into one data set and focusing on the relationship between δ¹³C_collagen and Δ_ca-co values (R² = 0.09, p < 0.0001). Our recalculation of their published data also yields significant (and stronger) relationships between δ¹³C_carbonate and Δ_ca-co values (R² = 0.34, R = 0.58, p < 0.0001 when calculated via enamel carbonate and dentin collagen in teeth; R² = 0.29, R = 0.54, p < 0.0001 when calculated via bone carbonate and bone collagen; and R² = 0.35, R = 0.59, p < 0.0001 when combined), though these results are weaker than those observed in carnivores (Supplementary Table S3). Within trophic-level difference in herbivores were attributed to physiology or environmental factors, to differences in methane production that result from grass vs. browse consumption, and to differences between ruminants and hindgut fermenters (Cerling and Harris, 1999; Clauss et al., 2020). However, as we noted with the springbok example generated from data by Codron et al. (2018)—its ‘environment’ (i.e., the composition of its diet, as browse or grass) is likely the driving factor of Δ_ca-co values in this bovid as all members of the species (Antidorcas marsupialis) share the same physiology.

Differences in canid and felid physiology are not likely the primary cause of Δ_ca-co variability. Canid and felid Δ_ca-co values are indistinguishable when feeding in similar environments (e.g., individuals of both P. concolor and C. latrans with similar δ¹³C_carbonate values have similar Δ_ca-co values in C₃ ecosystems, and the same is true for P. leo and L. pictus in C₄ ecosystems) but do vary with δ¹³C_carbonate values and both within and between C₃ or C₄ dominated environments. The fact that S. fatalis Δ_ca-co values are significantly lower than both canids at Rancho La Brea may stem from consuming prey that foraged in disparate environments (based on δ¹³C_carbonate values). In contrast, the Δ_ca-co values of the modern felid P. concolor and canid C. latrans that consume prey that forage in similar environments (as inferred from indistinguishable δ¹³C_carbonate values, p = 0.964) are indistinguishable from each other (p= 0.762). Thus, it does not appear that differences in canid and felid physiology are the primary cause of Δ_ca-co variability (see Figure 2). If physiological differences between canids and felids were a primary driver of Δ_ca-co variability, we would expect similar δ¹³C_carbonate values to yield disparate Δ_ca-co values in canids and felids, especially when feeding on similar prey. Accordingly, these data (Figures 2, 4) can help us rule out physiological differences between canids and felids as the primary driver. It therefore appears that the foraging habitats of prey have a substantial influence on Δ_ca-co values in carnivores, as documented here, and in herbivores as documented by Murphy et al. (2007) and Codron et al. (2018).

FIGURE 4

Figure 4. Relationships between δ¹³C_carbonate values and Δ_ca-co values for North American canid and felid specimens examined. Relationships between δ¹³C_carbonate values and Δ_ca-co values and in Canis latrans (A), Aenocyon dirus (B), and the felids (C; both Puma concolor and Smildon fatalis). Black lines and subsequent regression equations and R² values reflect the inclusion of all data per taxon. The dark-gray dashed line in panel B denotes the regression line and equations specific to A. dirus based on δ¹³C_collagen from bone collagen, while the solid black line includes δ¹³C_collagen data from tooth dentin from A. dirus from Locality 3,874. A light-gray dashed-line with a slope of 1 is present in the background of each panel for reference. *denotes significant relationships (p < 0.05).

Tissue and environmental/dietary drivers of collagen-carbonate δ¹³C spacing

Tissue specific differences, including processes governing how macronutrients such as amino acids and fats (including types and lengths of fatty acids) are incorporated into tooth enamel and bone collagen, may contribute to differences between δ¹³C_carbonate and δ¹³C_collagen and Δ_ca-co values. Earlier work by Ambrose and Norr (1993) established that δ¹³C_carbonate values reflect whole diet (carbohydrates, proteins, and fats), while δ¹³C_collagen reflects protein consumption. In carnivores, especially hypercarnivores, it is expected that carbohydrates would be a minor component of their diet and that the ‘whole diet’ most likely reflects proteins and fats, in contrast to primarily protein being routed into bone collagen (Ambrose and Norr, 1993). Further, as fat is depleted in ¹³C compared to either protein or carbohydrates (Ramsay and Hobson, 1991), one would expect δ¹³C_carbonate values to be more ¹³C depleted when consuming prey with a higher fat content (all else being equal; though research in this area is needed). Fats are ¹³C depleted compared to muscle to an astonishing degree, differing by as much as 7–8‰ in polar bears and ringed seals, with muscle and fat being even more ¹³C depleted than bone collagen (Tieszen et al., 1983; Ramsay and Hobson, 1991). The ¹³C depleted nature of fat is the primary reason why carnivores were thought to have lower Δ_ca-co values (due to the consumption of ¹³C depleted fat relative to other sources (Lee-Thorp et al., 1989; Clementz et al., 2009). However, an observation of Ambrose and Norr (1993) is often overlooked; when δ¹³C values of protein are higher than δ¹³C values of whole diet (regardless of the amount of protein in the diet; i.e., whether the diet consists of 6% protein or 77% protein) then Δ_ca-co values are low, when δ¹³C values of protein are lower than δ¹³C values of whole diet then Δ_ca-co values are high (again, regardless of the amount of protein in the diet, i.e., whether the diet consists of 6, 25, or 76% protein). Monoisotopic diets (where protein and whole diet are similar) yield intermediate Δ_ca-co values (i.e., 5.7+/−0.4‰, as compared to a range of 1.2 to 11.3‰; Ambrose and Norr, 1993). Thus, the isotopic composition of carbohydrates and fat relative to protein can influence Δ_ca-co values, and Δ_ca-co values are far from predictable based on trophic level or the amount of protein in one’s diet, alone.

Observed larger offsets in Δ_ca-co values in carnivores could stem from either increased consumption of more ¹³C enriched carbohydrates, the consumption of leaner prey, and/or different amounts of methane production. One possibility is that the canids consume more terrestrial resources that are not meat, and the higher Δ_ca-co values are a result of more non-meat sources. Alternatively, even limited non-protein sources with higher Δ_ca-co values (including fat) could result in higher Δ_ca-co values. Ambrose and Norr (1993) demonstrated that large Δ_ca-co values occur (i.e., Δ_ca-co values of 10.8+/−0.38, 10.8+/−0.36, and 11.3+/−0.40) when rats were fed protein that was significantly isotopically lighter than whole diet values (by ~10‰); in contrast, much smaller Δ_ca-co values occur (i.e., Δ_ca-co values of 1.2+/−0.10, 2.1+/ 0.24) when rats were fed protein that was isotopically heavier than whole diet values. As the canids have the highest Δ_ca-co values, as compared to the felids (from North America), increased C₄ whole diet values (which can include protein, carbohydrates, and fats) in canids could also cause this pattern. Aenocyon dirus has been interpreted as a hypercarnivore (Van Valkenburgh, 1991), making the source of the non-protein components less likely to be terrestrial plants and nuts, though even a small amount of supplemental foods that are isotopically heavier could drive this pattern (based on data from Ambrose and Norr, 1993) in addition to δ¹⁵N_collagen values providing only a ‘minimum estimate’ of trophic level (per Bocherens, 2015). Alternatively, the consumption of fat in prey from more open grassland ecosystems could also result in larger Δ_ca-co values that correlate with δ¹³C_carbonate values (per the discussion of Ambrose and Norr, 1993, above). Literature from the livestock industry demonstrates that free-ranging elk and bison can yield disparate fat and fatty acid compositions (Rule et al., 2002; e.g., elk have higher n-6 fatty acids than bison or beef cattle), which can have downstream impacts on the isotopic values observed in predators. Further, carnivores are known to increase fat oxidation when consuming a high-fat diet (Lester et al., 1999), which may also influence resulting Δ_ca-co values, though this has not yet been documented. While methane production is more pronounced in herbivores than carnivores (and most pronounced in grass-rich diets; Cerling and Harris, 1999), methane production could also play a role in carnivores if greater methane production occurs when digesting leaner prey (this could also yield higher Δ_ca-co values in carnivores that correlate with the foraging habitat of prey); however, this is not well understood.

Environmental differences, including water deficits may also contribute to isotopic offsets, as suggested by Murphy et al. (2007). Water deficits are positively correlated with Δ_ca-co values in kangaroos (Murphy et al., 2007); thus, as it gets wetter Δ_ca-co values decrease. While plants need adequate water to grow (zero rainfall results in no growth), excess rainfall beyond a certain threshold can reduce the nutritional quality of foliage (increased rainfall results in increased growth, and subsequently lower nitrogen per unit biomass, which is less nutritious; Austin and Vitousek, 1998). Similar effects are observed when growing plants under higher atmospheric CO₂ conditions. Plants grown under higher atmospheric CO₂ levels (which also have lower δ¹³C values when grown in growth chambers) also yield more growth (structural carbon, and often associated tissues like lignin and cellulose) while per unit biomass is subsequently less nutritious (nitrogen content is lower; Conroy, 1992, Barbehenn et al., 2004). All things being equal, herbivores that consume plants with relatively lower nitrogen/protein content may utilize protein differently and have different amounts of fat stores that are subsequently transferred to predators. More work is needed to clarify if and how environmental factors (e.g., precipitation, atmospheric CO₂) play a role in affecting herbivore and carnivore Δ_ca-co values.

Future research directions

It is not yet clear why Δ_ca-co and δ¹³C_carbonate from bone apatite are unrelated in African carnivores collectively (when combining canids and felids), both here and in Codron et al. (2018), in contrast to enamel apatite. However, there is clear evidence that bone apatite and enamel apatite tissues are not equal and vary quite dramatically in their mineralogy and crystal size (Wopenka and Pasteris, 2005). Bone apatite crystals typically lack the hydroxyl ion (OH⁻) and are about 10 to 100 times smaller than in enamel, both of which can affect carbonate concentration (with bone apatite having higher carbonate concentration and thus also greater solubility; Wopenka and Pasteris, 2005). Therefore, bone is well suited for being resorbed/re-precipitated and interacting with organic material, while enamel is well suited to resisting dissolution (Wopenka and Pasteris, 2005). While herbivores do exhibit positive relationships between δ¹³C_carbonate and Δ_ca-co in both bone apatite and enamel apatite, bone apatite has significantly lower Δ_ca-co values and δ¹³C_carbonate values than enamel (p < 0.0001, based on the analysis of data from Codron et al., 2018; mean Δ_ca-co in bone and enamel is 6.7+/-1.4‰ and 9.3+/-2.1‰, respectively; mean δ¹³C_carbonate values of-8.5+/-6.5‰, and-5.1+/-6.2‰, respectively). Further, our analysis of African carnivores also demonstrates no relationship between δ¹³C_carbonate in bone and Δ_ca-co and no relationship between δ¹³C_carbonate in dentin and Δ_ca-co when calculated from dentin carbonate and dentin collagen (Supplementary Table S3). The lack of a relationship between dentin δ¹³C_carbonate and dentin δ¹³C_collagen, similar to bone yet different from enamel, indicates that while differences in tissue routing may be at play, they likely do not stem from differences in routing due to ontogeny (e.g., using more amino acids for bone growth while young but not while older). Instead, these differences may have more to do with differences in the inorganic content of enamel versus both dentin and bone (and dentin vs. bone) and/or the mineralogy of these tissues (Wang and Cerling, 1994; Wopenka and Pasteris, 2005). Further, wolves (Canis lupus) document a significant positive relationship between δ¹³C_carbonate in bone and Δ_ca-co (based on data from Clementz et al., 2009), as do African wild dogs noted here, but in contrast to African lions (Supplementary Table S3). Thus, future work is needed to examine mechanisms behind these tissue specific discrepancies in Δ_ca-co values in carnivores, including differences between specific taxa.

More work is also needed to determine the precise mechanism responsible for the correlation between increased Δ_ca-co values and higher δ¹³C_carbonate values, and why this pattern is not apparent in bone apatite and dentin apatite of African carnivores (Codron et al., 2018, and noted here). Further, a re-evaluation of the potential drivers of relationships between herbivore Δ_ca-co and δ¹³C_carbonate values may be necessary. Codron et al. (2018) outlined two potential drivers of the positive relationship between Δ_ca-co values and δ¹³C_carbonate values in herbivores: environmental differences in food sources; and physiological differences between hindgut fermenters and ruminants. As carnivores did not exhibit the above relationship in their data set (which was primarily composed of bone and not teeth), it was assumed that environmental factors were not at play, as it was expected that environmental influences would similarly affect relationships between both herbivore and carnivore Δ_ca-co values and δ¹³C_carbonate values. Considering data presented here, carnivores clearly demonstrate a significant positive relationship between Δ_ca-co values and δ¹³C_carbonate values in tooth enamel, but not bone apatite (weakly in some; see Clementz et al., 2009). Thus, macronutrient routing in these disparate tissues (biochemically and structurally) may also play a role. These drivers are not mutually exclusive and may both contribute to the isotopic disparities outlined here. Further, we clearly document that carnivores in both C₃ and C₄ ecosystems demonstrate positive relationships between Δ_ca-co and δ¹³C_carbonate values (Figure 2), yet the reasons for ecosystem-associated differences between these lines and their intercepts is unclear—though may stem from differences in rainfall and/or CO₂ over time (and effects on protein and fat content in plants and prey). Thus, until the precise mechanisms driving the highly predictable Δ_ca-co values are identified, the examination of multiple tissues and isotopes may reveal a more complete understanding of the dietary ecology of both extinct and extant mammals.

Advances and challenges of relevance to paleobiology and archeology

The Δ_ca-co spacing in extinct and extant carnivores is predictable based on δ¹³C_carbonate values—indicating that the foraging behavior of prey consumed by predators is the best predictor of Δ_ca-co spacing. This positive relationship negates the idea that within trophic-level Δ_ca-co spacing is constant in carnivores and calls into question our ability to interpret degrees of carnivory, omnivory, or herbivory from Δ_ca-co values, alone. In the case of carnivores, felids and canids have similar Δ_ca-co values when consuming similar prey today (in both C₃ and C₄ environments), while felids and canids only differ in the past when consuming prey that foraged in different habitats. Thus, physiological differences between canids and felids do not appear to be drivers of Δ_ca-co spacing. While either ontogenetic differences in diet or diagenesis have been invoked as potential reasons for differences in δ¹³C_carbonate and δ¹³C_collagen values, neither of these hypotheses have much support. These data allow us to rule out many hypotheses and indicates that Δ_ca-co spacing is likely driven by the foraging habitat of prey, environmental influences, and/or tissue specific differences that affect how macronutrients are allocated and incorporated into mineralized tissues.

More work is needed to better understand the drivers of increased Δ_ca-co spacing in individuals with higher δ¹³C_carbonate values and why these relationships are weaker or absent when comparing certain tissues. Further, not all tissues can be treated similarly. For example, bone apatite and enamel apatite were combined in Codron et al. (2018) as δ¹³C values were similar; however, the resulting Δ_ca-co values are significantly higher when enamel carbonate and bone collagen were used to calculate Δ_ca-co (p < 0.0001, y-intercepts/mean differences of 2.6‰; Figure 3). While sampling from multiple tissues from the same specimens can be costly and potentially more destructive, more work that examines these relationships in different ecosystems and through time is necessary to better understand if and how a taxon’s habitat contributes to nutrient routing to different tissues.

Our understanding of Rancho La Brea carnivores has benefited from sampling both enamel and bone collagen. Isotopes from different tissues are not “right” or “wrong” but rather provide different insights into the dietary ecology of carnivores. Nitrogen isotope ratios demonstrate that C. latrans likely consumed less meat than either A. dirus or S. fatalis. Stable carbon isotope values measured in bone collagen from Coltrain et al. (2004), Fuller et al. (2020), and here indicate that the protein component of Rancho La Brea carnivore diets are similar, while carbon isotopes ratios from carbonate enamel indicate stark differences in the whole diet of S. fatalis as compared to A. dirus (DeSantis et al., 2019, 2020, and here) and C. latrans (at Pits 61/67). Differences in the stable carbon isotope signatures from whole diet indicate that A. dirus and S. fatalis are not consuming similar proportions of the same prey species and are instead consuming prey with disparate isotopic signatures—suggesting that these prey may forage in more open grasslands and more wooded areas, respectively; and/or may exhibit differences in fat content which can influence whole diet. As C₃ grasses have higher amounts of protein than C₄ grasses (Barbehenn et al., 2004); protein in prey-species may be biased toward C₃ protein sources, which may be responsible for similar carbon isotope values of bone collagen in herbivores and carnivores. The dominance of protein from C₃ plants can be further amplified in drier ecosystems, where C₃ plants can have even more protein per unit biomass than in higher rainfall regions (Austin and Vitousek, 1998). Fat is an important component of hypercarnivore diets (e.g., polar bears; Ramsay and Hobson, 1991), and many hypercarnivores preferentially consume fat (Nowak and Walker, 1999). However, the isotopic source of fat is primarily reflected in the carbonate portion of tooth enamel, dentin, or bone, but not in collagen (Ambrose and Norr, 1993). Thus, whole diet differences are apparent between S. fatalis and A. dirus (DeSantis et al., 2019, 2020, and here), and C. latrans from Pits 61/67. Protein component of one’s diet does not necessarily equal prey source, and we argue here that whole diet indicators (e.g., carbonate) are necessary to elucidating diets of extinct and extant carnivores. Our work here suggests potential drivers of the differences between stable carbon isotopes in enamel and collagen in Rancho La Brea taxa that include tissue and/or environmental differences (while challenging others including trophic level, ontogenetic dietary differences, physiological differences between canids and felids, and diagenesis), and outlines areas of future research that can bring further clarity to our understanding of the dietary ecology of predators and prey, today and in the past.

Data availability statement

The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author.

Ethics statement

Ethical review and approval was not required for the animal study because all of the animals studied were from museum collections. They were killed previously and not for the purpose of this study.

Author contributions

LD and RF conceptualized the study. All co-authors contributed to study design. LD, JS, JC, RF, and GT collected, prepared, and produced specimen/isotope data. LD analyzed stable isotope data. LD wrote the paper with significant editorial input from RF, JS, and JH, and additional feedback from all co-authors. All authors contributed to the article and approved the submitted version.

Funding

This work was supported by the National Science Foundation (1757545, 1758110, 1758116, 1758117, 1757236, and 1758108), Vanderbilt University, and the NY State Museum.

Acknowledgments

We thank L. Chiappe, P. Collins, P. Holroyd, E. Holt, and K. Fahy for access to materials, KF-D for discussions about stable isotopes and Rancho La Brea, J. Curtis for isotopic analysis, J. Olsson for the shadow drawings in Figures 1, 4, S. DeSantis for assistance with Figures 2, 4, and D. Fox, H. Bocherens, V. M. Bravo-Cuevas and additional reviewers for detailed feedback on earlier versions of this manuscript.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fevo.2022.1031383/full#supplementary-material

SUPPLEMENTARY TABLE S1 | Descriptive statistics of stable carbon and nitrogen isotope ratios from bone collagen and enamel carbonate of all taxa examined from North America.

SUPPLEMENTARY TABLE S2 | Descriptive statistics of stable carbon and nitrogen isotope ratios from bone, dentin, and enamel in African carnivores.

SUPPLEMENTARY TABLE S3 | Descriptive statistics of differences between carbon isotope values in carbonate and collagen (carbonate-collagen, Δca-co) in all North American and African carnivores examined in this study.

SUPPLEMENTARY TABLE S4 | Descriptive statistics of relationships between isotope values in carbonate and collagen and offset values in North American and African carnivores.

SUPPLEMENTARY TABLE S5 | Descriptive statistics of relationships between isotope ratios in carbonate and collagen from enamel, dentin, and bone in African carnivores from C4 ecosystems (n = 17).

SUPPLEMENTAL TABLE 1 | Stable isotope values and associated calculations for all individual samples (modern and fossil) from Rancho La Brea, southern California, and throughout Africa.

SUPPLEMENTAL TABLE 2 | Stable isotope data and associated calculations for all tissues (enamel, dentin, bone) from individual samples (modern) from Africa, housed in the American Museum of Natural History (AMNH).

References

Ambrose, S. H., Butler, B. M., Hanson, D. B., Hunter-Anderson, R. L., and Krueger, H. W. (1997). Stable isotopic analysis of human diet in the Marianas archipelago, Western Pacific. Am. J. Phys. Anthropol. 104, 343–361. doi: 10.1002/(SICI)1096-8644(199711)104:3<343::AID-AJPA5>3.0.CO;2-W

PubMed Abstract | CrossRef Full Text | Google Scholar

Ambrose, S. H., and Norr, L. (1993). “Experimental evidence for the relationship of the carbon isotope ratios of whole diet and dietary protein to those of bone collagen and carbonate,” in Prehistoric human bone: Archaeology at the molecular level. eds. J. B. Lambert and G. Grupe (Heidelberg, German: Springer), 1–37.

Google Scholar

Auerswald, K., Wittmer, M. H. O. M., Zazzo, A., Schaufele, R., and Schnyder, H. (2010). Biases in the analyses if stable isotope discrimination in food webs. J. Appl. Ecol. 47, 936–941. doi: 10.1111/j.1365-2664.2009.01764.x

CrossRef Full Text | Google Scholar

Austin, A. T., and Vitousek, P. M. (1998). Nutrient dynamics on a precipitation gradient in Hawai'i. Oecologia 113, 519–529. doi: 10.1007/s004420050405

PubMed Abstract | CrossRef Full Text | Google Scholar

Barbehenn, R. V., Chen, Z., Karowe, D. N., and Spickard, A. (2004). C₃ grasses have higher nutritional quality than C₄ grasses under ambient and elevated atmospheric CO₂. Glob. Chang. Biol. 10, 1565–1575. doi: 10.1111/j.1365-2486.2004.00833.x

CrossRef Full Text | Google Scholar

Barnosky, A. D., Koch, P. L., Feranec, R. S., Wing, S. L., and Shabel, A. B. (2004). Assessing the causes of late Pleistocene extinctions on the continents. Science 306, 70–75. doi: 10.1126/science.1101476

PubMed Abstract | CrossRef Full Text | Google Scholar

Bekoff, M., and Wells, M. C. (1982). Behavioral ecology of coyotes: social organization, rearing patterns, space use, and resource defense. Ethology 60, 281–305.

Google Scholar

Bocherens, H. (2015). Isotopic tracking of large carnivore palaeoecology in the mammoth steppe. Quat. Sci. Rev. 117, 42–71. doi: 10.1016/j.quascirev.2015.03.018

CrossRef Full Text | Google Scholar

Bocherens, H., Cotte, M., Bonini, R. A., Straccia, P., Scian, D., Soibelzon, L., et al. (2017). Isotopic insight on paleodiet of extinct Pleistocene megafaunal Xenarthrans from Argentina. Gondwana Res. 48, 7–14. doi: 10.1016/j.gr.2017.04.003

CrossRef Full Text | Google Scholar

Bocherens, H., Fizet, M., and Mariotti, A. (1994). Diet, physiology and ecology of fossil mammals as inferred from stable carbon and nitrogen isotope biogeochemistry: implications for Pleistocene bears. Palaeogeogr. Palaeoclimatol. Palaeoecol. 107, 213–225. doi: 10.1016/0031-0182(94)90095-7

CrossRef Full Text | Google Scholar

Carbone, C., Maddox, T., Funston, P. J., Mills, M. G., Grether, G. F., and Van Valkenburgh, B. (2009). Parallels between playbacks and Pleistocene tar seeps suggest sociality in an extinct sabretooth cat, Smilodon. Biol. Lett. 5, 81–85. doi: 10.1098/rsbl.2008.0526

PubMed Abstract | CrossRef Full Text | Google Scholar

Caut, S., Angulo, E., Courchamp, F., and Figuerola, J. (2010). Trophic experiments to estimate isotope discrimination factors. J. Appl. Ecol. 47, 948–954. doi: 10.1111/j.1365-2664.2010.01832.x

CrossRef Full Text | Google Scholar

Cerling, T. E., and Harris, J. M. (1999). Carbon isotope fractionation between diet and bioapatite in ungulate mammals and implications for ecological and paleoecological studies. Oecologia 120, 347–363. doi: 10.1007/s004420050868

PubMed Abstract | CrossRef Full Text | Google Scholar

Cerling, T. E., Harris, J. M., MacFadden, B. J., Leakey, M. G., Quade, J., Eisenmann, V., et al. (1997). Global vegetation change through the Miocene/Pliocene boundary. Nature 389, 153–158. doi: 10.1038/38229

CrossRef Full Text | Google Scholar

Clauss, M., Dittmann, M. T., Vendl, C., Hagen, K. B., Frei, S., Ortmann, S., et al. (2020). Comparative methane production in mammalian herbivores. Animal 14, s113–s123. doi: 10.1017/S1751731119003161

PubMed Abstract | CrossRef Full Text | Google Scholar

Clementz, M. T., Fox-Dobbs, K., Wheatley, P. V., Koch, P. L., and Doak, D. F. (2009). Revisiting old bones: coupled carbon isotope analysis of bioapatite and collagen as an ecological and palaeoecological tool. Geol. J. 44, 605–620. doi: 10.1002/gj.1173

CrossRef Full Text | Google Scholar

Clementz, M. T., Koch, P. L., and Beck, C. A. (2007). Diet induced differences in carbon isotope fractionation between sirenians and terrestrial ungulates. Mar. Biol. 151, 1773–1784. doi: 10.1007/s00227-007-0616-1

CrossRef Full Text | Google Scholar

Codron, D., Clauss, M., Codron, J., and Tütken, T. (2018). Within trophic level shifts in collagen–carbonate stable carbon isotope spacing are propagated by diet and digestive physiology in large mammal herbivores. Ecol. Evol. 8, 3983–3995. doi: 10.1002/ece3.3786

PubMed Abstract | CrossRef Full Text | Google Scholar

Collins, M. J., Nielsen-Marsh, C. M., Hiller, J., Smith, C. I., Roberts, J. P., Prigodich, R. V., et al. (2002). The survival of organic matter in bone: a review. Archaeometry 44, 383–394. doi: 10.1111/1475-4754.t01-1-00071

CrossRef Full Text | Google Scholar

Coltrain, J. B., Harris, J. M., Cerling, T. E., Ehleringer, J. R., Dearing, M.-D., Ward, J., et al. (2004). Rancho La Brea stable isotope biogeochemistry and its implications for the paleoecology of the late Pleistocene coastal southern California. Palaeogeogr. Palaeoclimatol. Palaeoecol. 205, 199–219. doi: 10.1016/j.palaeo.2003.12.008

CrossRef Full Text | Google Scholar

Conroy, J. P. (1992). Influence of elevated atmospheric CO2 concentrations on plant nutrition. Aust. J. Bot. 40, 445–456. doi: 10.1071/BT9920445

CrossRef Full Text | Google Scholar

Coplen, T. B. (1994). Reporting of stable hydrogen carbon and oxygen isotopic abundances. Pure Appl. Chem. 66, 273–276. doi: 10.1351/pac199466020273

CrossRef Full Text | Google Scholar

DeNiro, M. J., and Epstein, S. (1978). Influence of diet on the distribution of carbon isotopes in animals. Geochim. Cosmochim. Acta 42, 495–506. doi: 10.1016/0016-7037(78)90199-0

CrossRef Full Text | Google Scholar

DeSantis, L. R. G., Crites, J. M., Feranec, R. S., Fox-Dobbs, K., Farrell, A. B., Harris, J. M., et al. (2019). Causes and consequences of Pleistocene megafaunal extinctions as revealed from rancho La Brea mammals. Curr. Biol. 29, 2488–2495.e2. doi: 10.1016/j.cub.2019.06.059

PubMed Abstract | CrossRef Full Text | Google Scholar

DeSantis, L. R., Feranec, R. S., Fox-Dobbs, K., Harris, J. M., Cerling, T. E., Crites, J. M., Farrell, A. B., and Takeuchi, G. T.. 2020. Reply to Van Valkenburgh, et al. Curr. Biol. 30:R151–R152, doi: 10.1016/j.cub.2020.01.011

PubMed Abstract | CrossRef Full Text | Google Scholar

DeSantis, L. R. G., Feranec, R. S., and MacFadden, B. J. (2009). Effects of global warming on ancient mammalian communities and their environments. PLoS One 4:e5750. doi: 10.1371/journal.pone.0005750

PubMed Abstract | CrossRef Full Text | Google Scholar

DeSantis, L. R. G., Schubert, B. W., Scott, J. R., and Ungar, P. S. (2012). Implications of diet for the extinction of saber-toothed cats and American lions. PLoS One 7:e5245. doi: 10.1371/journal.pone.0052453

PubMed Abstract | CrossRef Full Text | Google Scholar

Elliot, J. P., Cowan, I. M., and Holling, C. S. (1977). Prey capture by the African lion. Can. J. Zool. 55, 1811–1828. doi: 10.1139/z77-235

CrossRef Full Text | Google Scholar

Ewer, R. F. (1998). The carnivores Ithaca, NY: Cornell University Press.

Google Scholar

Feranec, R. S. (2004). Isotopic evidence of saber-tooth development, growth rate, and diet from the adult canine of Smilodon fatalis from rancho La Brea. Palaeogeogr. Palaeoclimatol. Palaeoecol. 206, 303–310. doi: 10.1016/j.palaeo.2004.01.009

CrossRef Full Text | Google Scholar

Feranec, R. S., and DeSantis, L. R. G. (2014). Understanding specifics in generalist diets of carnivorans by analyzing stable carbon isotope values in Pleistocene mammals of Florida. Paleobiology 40, 477–493. doi: 10.1666/13055

CrossRef Full Text | Google Scholar

Feranec, R. S., Hadly, E. A., and Paytan, A. (2009). Stable isotopes reveal seasonal competition for resources between late Pleistocene bison (Bison) and horse (Equus) from rancho La Brea, southern California. Palaeogeogr. Palaeoclimatol. Palaeoecol. 271, 153–160. doi: 10.1016/j.palaeo.2008.10.005

CrossRef Full Text | Google Scholar

Finucane, B., Agurto, P. M., and Isbell, W. H. (2006). Human and animal diet at Conchopata, Peru: stable isotope evidence for maize agriculture and animal management practices during the middle horizon. J. Archaeol. Sci. 33, 1766–1776. doi: 10.1016/j.jas.2006.03.012

CrossRef Full Text | Google Scholar

Friedman, I., and O'Neil, J. R. (1977). Compilation of stable isotope fractionation factors of geochemical interest, vol. 440 US Government Printing Office.

Google Scholar

Fuller, B. T., Fahrni, S. M., Harris, J. M., Farrell, A. B., Coltrain, J. B., Gerhart, L. M., et al. (2014). Ultrafiltration for asphalt removal from bone collagen for radiocarbon dating and isotopic analysis of Pleistocene fauna at the tar pits of rancho La Brea, Los Angeles, California. Quat. Geochronol. 22, 85–98. doi: 10.1016/j.quageo.2014.03.002

CrossRef Full Text | Google Scholar

Fuller, B. T., Fuller, J. L., Harris, D. A., and Hedges, R. E. M. (2006). Detection of breastfeeding and weaning in modern human infants with carbon and nitrogen stable isotope ratios. Am. J. Phys. Anthropol. 129, 279–293. doi: 10.1002/ajpa.20249

PubMed Abstract | CrossRef Full Text | Google Scholar

Fuller, B. T., Southon, J. R., Fahrni, S. M., Farrell, A. B., Takeuchi, G. T., Nehlich, O., et al. (2020). Pleistocene paleoecology and feeding behavior of terrestrial vertebrates recorded in a pre-LGM asphaltic deposit at rancho La Brea, California. Palaeogeogr. Palaeoclimatol. Palaeoecol. 537:109383. doi: 10.1016/j.palaeo.2019.109383

CrossRef Full Text | Google Scholar

Haas, S. K., Hayssen, B., and Krausman, P. R. (2005). Panthera leo. Mamm. Species 762, 1–11. doi: 10.1644/1545-1410(2005)762[0001:PL]2.0.CO;2

CrossRef Full Text | Google Scholar

Hedges, R. E. (2003). On bone collagen—apatite-carbonate isotopic relationships. Int. J. Osteoarchaeol. 13, 66–79. doi: 10.1002/oa.660

CrossRef Full Text | Google Scholar

Hedges, R. E., Clement, J. G., Thomas, C. D., and O’Connell, T. C. (2007). Collagen turnover in the adult femoral mid-shaft: modelled from anthropogenic radiocarbon tracer measurements. Am. J. Phys. Anthropol. 133, 808–816. doi: 10.1002/ajpa.20598

PubMed Abstract | CrossRef Full Text | Google Scholar

Helfman, P. M., and Bada, J. L. (1976). Aspartic acid racemisation in dentine as a measure of ageing. Nature 262, 279–281. doi: 10.1038/262279b0

PubMed Abstract | CrossRef Full Text | Google Scholar

Jenkins, S. G., Partridge, S. T., Stephenson, T. R., Farley, S. D., and Robbins, C. T. (2001). Nitrogen and carbon isotope fractionation between mothers, neonates, and nursing offspring. Oecologia 129, 336–341. doi: 10.1007/s004420100755

PubMed Abstract | CrossRef Full Text | Google Scholar

Jones, D. B., and DeSantis, L. R. G. (2017). Dietary ecology of ungulates from the La Brea tar pits in southern California: a multi-proxy approach. Palaeogeogr. Palaeoclimatol. Palaeoecol. 466, 110–127. doi: 10.1016/j.palaeo.2016.11.019

CrossRef Full Text | Google Scholar

Keeling, C. D. (1979). The Suess effect: 13Carbon-14Carbon interrelations. Environ. Int. 2, 229–300. doi: 10.1016/0160-4120(79)90005-9

CrossRef Full Text | Google Scholar

Keeling, C. D., Mook, W. G., and Tans, P. P. (1979). Recent trends in the 13C/12C ratio of atmospheric carbon dioxide. Nature 277, 121–123. doi: 10.1038/277121a0

CrossRef Full Text | Google Scholar

Kellner, C. M., and Schoeninger, M. J. (2007). A simple carbon isotope model for reconstructing prehistoric human diet. Am. J. Phys. Anthropol. 133, 1112–1127. doi: 10.1002/ajpa.20618

PubMed Abstract | CrossRef Full Text | Google Scholar

Kendall, C., Eriksen, A. M., Kontopoulos, I., Collins, M. J., and Gordon, T.-W. (2018). Diagenesis of archaeological bone and tooth. Palaeogeogr. Palaeoclimatol. Palaeoecol. 491, 21–37. doi: 10.1016/j.palaeo.2017.11.041

CrossRef Full Text | Google Scholar

Koch, P. L. (2007). “Isotopic study of the biology of modern and fossil vertebrates,” in Stable isotopes in ecology and environmental science. eds. R. Michener and K. Lajtha. Second ed (Malden, MA: Blackwell), 99–154.

Google Scholar

Koch, P. L., Tuross, N., and Fogel, M. L. (1997). The effects of sample treatment and diagenesis on the isotopic integrity of carbonate in biogenic hydroxylapatite. J. Archaeol. Sci. 24, 417–429. doi: 10.1006/jasc.1996.0126

CrossRef Full Text | Google Scholar

Krueger, H. W., and Sullivan, C. H. (1984). “Models for carbon isotope fractionation between diet and bone,” in Stable isotopes in nutrition. ACS symposium series 258. eds. J. F. Turnland and P. E. Johnson (Washington D.C: American Chemical Society), 205–222.

Google Scholar

Lee-Thorp, J. A. (2008). On isotopes and old bones. Archaeometry 50, 925–950. doi: 10.1111/j.1475-4754.2008.00441.x

CrossRef Full Text | Google Scholar

Lee-Thorp, J. A., Sealy, J. C., and van der Merwe, N. J. (1989). Stable carbon isotope ratio differences between bone collagen and bone apatite, and their relationship to diet. J. Archaeol. Sci. 16, 585–599. doi: 10.1016/0305-4403(89)90024-1

CrossRef Full Text | Google Scholar

Lester, T., Czarnecki-Maulden, G., and Lewis, D. (1999). Cats increase fatty acid oxidation when isocalorically fed meat-based diets with increasing fat content. Regul. Integr. Physiol. 277, R878–R886. doi: 10.1152/ajpregu.1999.277.3.R878

PubMed Abstract | CrossRef Full Text | Google Scholar

Levin, N. E., Cerling, T. E., Passey, B. H., Harris, J. M., and Ehleringer, J. R. (2006). A stable isotope aridity index for terrestrial environments. Proc. Natl. Acad. Sci. 103, 11201–11205. doi: 10.1073/pnas.0604719103

PubMed Abstract | CrossRef Full Text | Google Scholar

MacFadden, B. J., DeSantis, L. R. G., Hochstein, J. L., and Kamenov, G. D. (2010). Physical properties, geochemistry, and diagenesis of xenarthran teeth: prospects for interpreting the paleoecology of extinct species. Palaeogeogr. Palaeoclimatol. Palaeoecol. 291, 180–189. doi: 10.1016/j.palaeo.2010.02.021

CrossRef Full Text | Google Scholar

Mariotti, A. (1983). Atmospheric nitrogen is a reliable standard for natural 15N measurements. Nature 303, 685–687. doi: 10.1038/303685a0

CrossRef Full Text | Google Scholar

McHorse, B. K., Orcutt, J. D., and Davis, E. B. (2012). The carnivoran fauna of rancho La Brea: average or aberrant? Palaeogeogr. Palaeoclimatol. Palaeoecol. 329-330, 118–123. doi: 10.1016/j.palaeo.2012.02.022

CrossRef Full Text | Google Scholar

Meachen, J. A., and Samuels, J. X. (2012). Evolution in coyotes (Canis latrans) in response to the megafaunal extinctions. Proc. Natl. Acad. Sci. U.S.A. 109, 4191–4196. doi: 10.1073/pnas.1113788109

PubMed Abstract | CrossRef Full Text | Google Scholar

Mech, L. D. (1974). Canis lupus. Mamm. Species 37, 1–6. doi: 10.2307/3503924

CrossRef Full Text | Google Scholar

Merriam, J. C., and Stock, C.. 1932. The Felidae of rancho la Brea. No. 422. Carnegie Institute, Washington.

Google Scholar

Murphy, B. P., Bowman, D. M. J. S., and Gagan, M. K. (2007). Sources of carbon isotope variation in kangaroo bone collagen and tooth enamel. Geochim. Cosmochim. Acta 71, 3847–3858. doi: 10.1016/j.gca.2007.05.012

CrossRef Full Text | Google Scholar

Nowak, R. M., and Walker, E. P. (1999). Walker's mammals of the world Baltimore, MD: John Hopkins University Press.

Google Scholar

O’Keefe, F. R., Fet, E. V., and Harris, J. M. (2009). Compilation, calibration, and synthesis of faunal and floral radiocarbon dates, Rancho La Brea, California. Cont. Sci. 518, 1–16. doi: 10.5962/p.226783

CrossRef Full Text | Google Scholar

O’Connell, T. C., and Hedges, R. E. M. (2017). Chicken and egg: testing the carbon isotopic effects of carnivory and herbivory. Archaeometry 59, 302–315. doi: 10.1111/arcm.12253

CrossRef Full Text | Google Scholar

Passey, B. H., Robinson, T. F., Ayliffe, L. K., Cerling, T. E., Sponheimer, M., Dearing, M. D., et al. (2005). Carbon isotope fractionation between diet, breath CO2, and bioapatite in different mammals. J. Archaeol. Sci. 32, 1459–1470. doi: 10.1016/j.jas.2005.03.015

CrossRef Full Text | Google Scholar

Ramsay, M. A., and Hobson, K. A. (1991). Polar bears make little use of terrestrial food webs: evidence from stable-carbon isotope analysis. Oecologia 86, 598–600. doi: 10.1007/BF00318328

PubMed Abstract | CrossRef Full Text | Google Scholar

Richards, M. P., Mays, S., and Fuller, B. T. (2002). Stable carbon and nitrogen isotope values of bone and teeth reflect weaning age at the Medieval Wharram Percy site, Yorkshire, UK. Am. J. Phys. Anthropol. 119, 205–210. doi: 10.1002/ajpa.10124

PubMed Abstract | CrossRef Full Text | Google Scholar

Rule, D. C., Broughton, K. S., Shellito, S. M., and Maiorano, G. (2002). Comparison of muscle fatty acid profiles and cholesterol concentrations of bison, beef cattle, elk, and chicken. J. Anim. Sci. 80, 1202–1211. doi: 10.2527/2002.8051202x

PubMed Abstract | CrossRef Full Text | Google Scholar

Samuels, J. X., Meachen, J. A., and Sakai, S. A. (2013). Postcranial morphology and the locomotor habits of living and extinct carnivorans. J. Morphol. 274, 121–146. doi: 10.1002/jmor.20077

PubMed Abstract | CrossRef Full Text | Google Scholar

Schaller, G. B. 1972. The Serengeti lion, a study of predator-prey relationships. Chicago; University of Chicago Press.

Google Scholar

Schoeninger, M. J., DeNiro, M. J., and Tauber, H. (1983). Stable nitrogen isotope ratios of bone collagen reflect marine and terrestrial components of prehistoric human diet. Science 220, 1381–1383. doi: 10.1126/science.6344217

PubMed Abstract | CrossRef Full Text | Google Scholar

Secord, R., Bloch, J. I., Chester, S. G., Boyer, D. M., Wood, A. R., Wing, S. L., et al. (2012). Evolution of the earliest horses driven by climate change in the Paleocene-Eocene thermal maximum. Science 335, 959–962. doi: 10.1126/science.1213859

PubMed Abstract | CrossRef Full Text | Google Scholar

Smith, F. A., Elliott Smith, E. A., Villaseñor, A., Tomé, C. P., Lyons, S. K., and Newsome, S. D. (2022). Late Pleistocene megafauna extinction leads to missing pieces of ecological space in a north American mammal community. Proc. Natl. Acad. Sci. 119:e2115015119. doi: 10.1073/pnas.2115015119

PubMed Abstract | CrossRef Full Text | Google Scholar

Stock, C., and Harris, J. M.. 1992. Rancho La Brea: A record of Pleistocene life in California Los Angeles: Natural History Museum of Los Angeles. Natural History Museum of Los Angeles County, Los Angeles.

Google Scholar

Teruel, J. D., Alcolea, A., Hernández, A., and Ruiz, A. J. O. (2015). Comparison of chemical composition of enamel and dentine in human, bovine, porcine and ovine teeth. Arch. Oral Biol. 60, 768–775. doi: 10.1016/j.archoralbio.2015.01.014

PubMed Abstract | CrossRef Full Text | Google Scholar

Tieszen, L. L., Boutton, T. W., Tesdahl, K. G., and Slade, N. A. (1983). Fractionation and turnover of stable carbon isotopes in animal tissues: implications for δ¹³C analysis of diet. Oecologia 57, 32–37. doi: 10.1007/BF00379558

PubMed Abstract | CrossRef Full Text | Google Scholar

Tsutaya, T., and Yoneda, M. (2015). Reconstruction of breastfeeding and weaning practices using stable isotope and trace element analyses: a review. Am. J. Phys. Anthropol. 156, 2–21. doi: 10.1002/ajpa.22657

PubMed Abstract | CrossRef Full Text | Google Scholar

Tung, T. A., Dillehay, T. D., Feranec, R. S., and DeSantis, L. R. G. (2020). Early specialized maritime and maize economies of the north coast of Peru. Proc. Natl. Acad. Sci. U.S.A. 117, 32308–32319. doi: 10.1073/pnas.2009121117

PubMed Abstract | CrossRef Full Text | Google Scholar

Tung, T. A., Miller, M., DeSantis, L., Sharp, E. A., and Kelly, J. (2016). “Patterns of violence and diet among children during a time of imperial decline and climate change in the ancient Peruvian Andes,” in The archaeology of food and warfare, eds. E. Amber, M. VanDerwarker and G. D. Wilson (Cham, Switzerland: Springer), 193–228.

Google Scholar

Van Valkenburgh, B. (1991). Iterative evolution of hypercarnivory in canids (Mammalia: Carnivora): evolutionary interactions among sympatric predators. Paleobiology 17, 340–362. doi: 10.1017/S0094837300010691

CrossRef Full Text | Google Scholar

Van Valkenburgh, B., Clementz, M. T., and Ben-David, M. (2020). Problems with inferring a lack of competition between rancho La Brea dire wolves and sabertooth cats based on dental enamel. Curr. Biol. 30, R149–R150. doi: 10.1016/j.cub.2020.01.009

PubMed Abstract | CrossRef Full Text | Google Scholar

Van Valkenburgh, B., and Hertel, F. (1993). Tough times at La Brea: tooth breakage in large carnivores of the late Pleistocene. Science 261, 456–459. doi: 10.1126/science.261.5120.456

PubMed Abstract | CrossRef Full Text | Google Scholar

Wang, Y., and Cerling, T. E. (1994). A model of fossil tooth and bone diagenesis: implications for paleodiet reconstruction from stable isotopes. Palaeogeogr. Palaeoclimatol. Palaeoecol. 107, 281–289. doi: 10.1016/0031-0182(94)90100-7

CrossRef Full Text | Google Scholar

Wopenka, B., and Pasteris, J. D. (2005). A mineralogical perspective on the apatite in bone. Mater. Sci. Eng. C 25, 131–143. doi: 10.1016/j.msec.2005.01.008

CrossRef Full Text | Google Scholar

Wysocki, M. A., Feranec, R. S., Tseng, Z. J., and Bjornsson, C. S. (2015). Using a novel absolute ontogenetic age determination technique to calculate the timing of tooth eruption in the saber-toothed cat, Smilodon fatalis. PLoS One 10:e0129847. doi: 10.1371/journal.pone.0129847

PubMed Abstract | CrossRef Full Text | Google Scholar